24 Hours Remaining for the $MEGA Public Sale
Reminders:
→ Refunds will be automatic (Nov 5th if no allocation)
→ Non-sybil bidders are eligible to participate in the Mainnet Campaign
→ Allocations will be known 1 week after auction close
The future will be defined by applications which think bigger than our industry.
Global Token Exchange is accelerating beyond app wars to represent a new paradigm of financial products which leverage blockchains to just be better.
Congrats @GTE_XYZ.
Looking for things to do on testnet? Use fluffle[.]tools
→ Live apps
→ Upcoming NFT mints
→ Individual application guides
h/t @0x_ultra for the community tool
No weekly reports for the MegaETH testnet were necessary for the past few weeks because they were pretty uneventful. This week, however, we do have a weekly report, because we had an unplanned maintenance. Here’s a breakdown.
On Friday morning ET, our monitor system reported a sudden drop in the testnet’s usage. This almost always indicates a problem, because the usage has been at 900+ tps for quite a while, so we immediately started responding.
Within a few minutes, we identified that pending transactions in the transaction pool (the module that buffers incoming transactions before they get a chance to be executed by the EVM) were quickly piling up. This was abnormal because the chain’s usage at that point was nowhere close to the capacity, so queuing should be almost nonexistent. We tried to clear the transaction pool, but unsurprisingly, pending transactions immediately began to reappear.
Looking at the transactions stuck in the pool, we quickly realized that they were causing live locks. These transaction would get into the EVM fine (because their senders had enough balance to pay for gas and the gas price was set appropriately), but the EVM would then abort execution (because the transactions violated one of the rules enforced by the EVM), and the failed transactions would be put back to the pool waiting for another try which was bound to also fail. This loop got repeated and the EVM was busy trying and failing these transactions without doing useful work (processing other transactions).
Although temporary fixes exist (which we successfully used an as interim solution in another occasion), we wanted a more permanent solution. A patch was coded up and tested within 2 hours and deployed in the production system, which fixed the issue. Throughout the incident, block (EVM and mini block) production and RPC service continued normally. Because of the issue, fewer legitimate transaction than normal got into the blocks during the incident.
This incident showed us again that increasing the EVM’s gas limit is an extremely intricate process. A high-performance execution engine is apparently very important, but other components such as the transaction pool, the peer-to-peer network, and the RPC service are equally so. The past 2 months of the MegaETH’s testnet has been extremely valuable for the team, and we thank everyone who helped us test the chain. We look forward to more testing, breaking, and building – in the open!
NEW MONTH IS HERE : Just a reminder to keep interacting with👇🏻
1. @monad_xyz Testnet
2. @megaeth Testnet
3. @inconetwork Testnet
4. @union_build Testnet
5. @irys_xyz Testnet
JOIN TG FOR MORE ALPHA:
https://t.co/QEey1sn3ve